AI Sidebar Spoofing Attack: SquareX Uncovers Malicious Extensions that Impersonate AI Browser Sidebars

SquareX has released research detailing a new class of attack called the AI Sidebar Spoofing attack, which uses malicious browser extensions to impersonate trusted AI sidebar interfaces. The findings, published on 24 October 2025, describe how attackers create pixel-perfect replicas of legitimate AI sidebars to feed users AI-generated instructions that conceal harmful prompts, enabling credential theft, device hijacking and password exfiltration.

The report notes that these attacks affect AI browsers such as Comet and consumer browsers with AI features like Edge, Firefox and Safari, meaning organisations restricting AI browsers may still see users exposed on any browser with an AI sidebar.

It also highlights that the extensions require only basic browser permissions and can remain dormant, appearing legitimate until a malicious prompt triggers wrongdoing, underscoring the need for runtime dynamic analysis and granular browser-native guardrails to warn and block users.

“AI has become an essential tool for millions of users to learn new skills and complete tasks,” according to Vivek Ramachandran, Founder and CEO of SquareX, “but people blindly follow AI-generated instructions without the expertise to identify security risks.”